Criminological Prediction - An Introduction (From Prediction in Criminology, P 2-33, 1985, David P Farrington and Roger Tarling, ed. - See NCJ-99006)

• NCJ Number: 99007
• Author(s): D P Farrington; R Tarling
• Date Published: 1985
• Length: 32 pages
• Sale Source: State University of New York Press
• Publication Type: Report (Study/Research)
• Language: English
• Country: United States of America

Annotation

This chapter examines prediction methods and issues in criminological research, with a focus on multivariate prediction and the selection, combination, and measurement of predictor variables.

Abstract

Specific application of prediction methods are presented in summaries of studies of dangerousness, delinquency, evaluation of penal treatments and parole, and police and correctional decisionmaking. Some of the methods used include scoring systems, scale, logistic regression, multivariate analyses of personality and other psychometric measures, Burgess-type point scores, and comparative analyses. Forecasting (e.g., future crime rates, resource allocation, criminal justice system processing) using mathematical and computer models also is examined. Basic issues in predictive research are discussed, including the availability of data, selection of predictor variables and measures, combining predictors, selecting criterion variables, measuring predictive efficiency, and the determination of shrinkage through comparisons of data from construction and validation samples. Emphasis is placed on the need for a theoretical foundation in the selection of predictors, criteria, and methods of combining predictors into a prediction instrument. Ethical issues in prediction research and areas for further research also are considered. A total of 125 references are listed.